Image Classification on ImageNet 1k (val) (Accuracy & Inference Efficiency)
154FLOPs (M)HALP
Evaluation results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| HALPModel=MobileNet-V1, Target=40%2022.10 | 154 | 67.2 | 87.32 | 8,293 | 2.43 | |
| HALPModel=MobileNet-V1, Target=42%2022.10 | 171 | 68.3 | 88.08 | 7,940 | 2.32 | |
| HALPModel=MobileNet-V2, Target=60%2022.10 | 183 | 70.42 | 89.75 | 5,668 | 1.84 | |
| HALPModel=MobileNet-V2, Target=65%2022.10 | 218 | 71.41 | 90.08 | 5,003 | 1.62 | |
| HALPModel=MobileNet-V2, Target=70%2022.10 | 227 | 71.88 | 90.39 | 4,478 | 1.45 | |
| HALPModel=MobileNet-V1, Target=50%2022.10 | 237 | 69.79 | 89.08 | 6,887 | 2.02 | |
| HALPModel=MobileNet-V2, Target=75%2022.10 | 249 | 72.16 | 90.44 | 4,109 | 1.33 | |
| HALPModel=MobileNet-V2, Target=90%2022.10 | 273 | 72.45 | 90.68 | 3,443 | 1.12 | |
| HALPModel=MobileNet-V2, Target=95%2022.10 | 281 | 72.55 | 90.79 | 3,265 | 1.06 | |
| HALPModel=MobileNet-V1, Target=60%2022.10 | 297 | 71.31 | 90.05 | 5,754 | 1.68 | |
| No pruningModel=MobileNet-V22022.10 | 301 | 72.1 | 90.6 | 3,080 | 1 | |
| HALPModel=MobileNet-V1, Target=70%2022.10 | 360 | 71.78 | 90.39 | 4,870 | 1.43 | |
| HALPModel=MobileNet-V1, Target=80%2022.10 | 416 | 72.52 | 90.78 | 4,167 | 1.22 | |
| HALPModel=MobileNet-V1, Target=90%2022.10 | 507 | 72.95 | 91.02 | 3,765 | 1.1 | |
| No pruningModel=MobileNet-V12022.10 | 569 | 72.64 | 90.88 | 3,415 | 1 |